Brushless Drag Bike!

[Arlo1]

Ok so if we compare to max torque in the 2 page of this thread at 3253 ft/lbs We can see a big advantage With an electric built with a 750000 watt goal in mind as rpm goes down torque goes up!
So with a 100% efficiecnt motor at all rpms we would have 1000 hp x 5252 / 100 = 52,520 ft lbs torque at 100 rpm and scaling down to 1000 ft/lbs at 5252! So if eveything is 100% efficient and it was a 1:1 drive those would be rear wheel numbers! And at 2170 rpm it will be 200 mph almost current nitro harley trap speed so 1000 hp x 5252 /2170 rpm = 2420 ft/lbs. Where as the Nitro guys never get more then 3253 ft/lbs in first gear and 2374 ft/lbs in second gear!
Now ..... How do we calculate efficiency losses for a motor we have not yet invented?

 

[bigmoose]

Your drive system will likely optimize around 800V to 1,000V, yep that's 1 KiloVolt. Drops the amps an order of magnitude from the 100 V bus. This is kill me now AND kill me later territory...

One could design around Hitachi MBN1500E33E2 IGBT modules. Data sheet: http://www.hitachi.co.jp/products/power/pse/images/pdf/igbt/MBN1500E33E2.pdf Those modules are around 5 x 8 inches each! They are rated at 3.3KiloVolts, 1,500 Amps. With a 3,000 Amp surge for 1mS, and a great reverse bias safe operating area. Rise and fall times are 2 uSec with turn off/turn on times around 3 or 4 uSec. So you are likely to run an audible PWM base frequency.

Snubbing and low inductance laminated buss bars are in the mix, as is about 500 hours of computer modeling of the layout. If you stay with 100V you might as well say you are switching a ga-billion-amps/second... handling your switching spikes will take 2 PdD's... and an infi-tacitor or two!

Just thinking of plugging together a 1 KiloVolt LiPo pack is making my hands sweat...

 

[Kingfish]

Thanks Arlo for being flexible! It is always best to use components that are designed for the task at hand, and I am a firm believer that less parts generally equals better reliability. Using more than 3-phases is intriguing and deserves investigation. Reviewing what the competition has applied also is beneficial. Efficiency calc momentarily.

BigMoose:
Yeah, my prelim calc had 2068 V @ 360.64 Amps, which works out as 1194 V / phase; that’s challenging to say the least for motor construction, though I like the amps. If you have 8 stators in parallel that’s 45 Amps each.

Read, read, and more reading:
Might I suggest that we all get on the same page – literally with the same reference books so we can share insights and reflect upon the possibilities together. I have been using:

Axial Flux Permanent Magnet Brushless Machines by Giera/Wang/Kamper; Publisher: Springer, 2nd Ed. © 2008. ISBN 987-1-4020-6993-2, and it comes with a CD of all the example formulas written for MathCAD (should anyone have access). It wasn’t a cheap book, but I read from it more than the Bible, Koran, or Buddha teachings… though arguable Buddha was pretty cool. Actually I never read the Bible or Koran so you’re in luck. However, in The Book of AF (see, I’m already turning it into a Gospel) there are several examples of motors and generators in the multi-kW range spinning at 3000 rpm and greater. No single example fits our task – but there’s enough there to guide us along with a boatload of math.

Wire:
Given

• F = IL x B
  Ï„ = 2rF = 2rBIL
  B = 0.5 T (this value could be higher depending on the magnets and air gap)
  r = 15.5 inches / 0.3937 meters
  I = 360.64 A
  Ï„ = 2189 Nm

Solve for L:

• L = Ï„ / 2rBI -> 2189 / (2 * 0. 3937 * 0.5 * 360.64) = 15.42 m

Heat by the System:
The heat generated by the system would be as follows...
It turns out that 0000 AWG can handle a maximum of 380 amps, and the resistance is 0.16072 Ohms/km.

• R = 0.16072 ohms/km for 0000 AWG -> (0.16072 * 15.42) / 1000 = 2.478 ohms
  R = V / I
  P = I^2 * R = V^2 / R
  P = 361^2 * 2.478 = 323 kW!

The power of the Sun between your legs; get it while it’s hot!

For kicks, what if we raised the magnetic flux density to 1.0 Tesla?
L = Ï„ / 2rBI -> 2189 / (2 * 0. 3937 * 1 * 360.64) = 7.67 m (This is why we want strong magnets!)
R = 0.16072 ohms/km for 0000 AWG -> (0.16072 * 7.67) / 1000 = 1.233 ohms
P = 361^2 * 1.233 = 161 kW!

Efficiency:
Pe = (Pi – Pr) / Pi -> (746 – 161) / 746 = 78.4%

This tells me that 0000 AWG wire is too resistant, and that we need better conductors. Unfortunately 0000 AWG is as high as my reference will go. Perhaps bus-bars would work better here, or a warm hyper-conductor of sorts. Regardless I am optimistic we can drop resistance down and increase our efficiency. I mean we are still baby-stepping and not yet out of the crib.

Have hope. KF

 

[liveforphysics]

Lol. 0000 awg can handle 380amps when its trapped in an insulated conduit with 10 other wires going through a 150deg section of a gen, and loaded 24-7.

Those wire size charts are for hugging retards.
Remember, your x5 motor wires are 4 little 22awg wires... We inherently know they handle hundreds of amps for long enough to finish a quarter mile.



For a race EV, 0000awg would be fine for 5KA.

 

[liveforphysics]

Lol, some wire charts would say not to exceed 6 amps to the motor windings... Lol!

 

[bradstuff]

Here is a link to the "other electric bike" that has taken the 1/4 record from the Killacycle.
http://www.paradefloats.com/orange.html
Shawn Lawless is the owner of the bike. "Larry ‘Spiderman’ McBride rode the 363 volt, 4000 amp electric bike to a 7.469, 177 miles per hour quarter-mile pass shattering the existing world record of 7.82 seconds at 168mph. " The bike has a 13" diameter GE brushed motor and two Zilla controllers.

Just summited as a starting point.

 

[AussieJester]

Here is a link to the "other electric bike" that has taken the 1/4 record from the Killacycle.
http://www.paradefloats.com/orange.html
Shawn Lawless is the owner of the bike. "Larry ‘Spiderman’ McBride rode the 363 volt, 4000 amp electric bike to a 7.469, 177 miles per hour quarter-mile pass shattering the existing world record of 7.82 seconds at 168mph. " The bike has a 13" diameter GE brushed motor and two Zilla controllers.

Just summited as a starting point.

old news...

http://endless-sphere.com/forums/viewtopic.php?f=3&t=21886&hilit=american+chopper

KiM

 

[Arlo1]

Here is a link to the "other electric bike" that has taken the 1/4 record from the Killacycle.
http://www.paradefloats.com/orange.html
Shawn Lawless is the owner of the bike. "Larry ‘Spiderman’ McBride rode the 363 volt, 4000 amp electric bike to a 7.469, 177 miles per hour quarter-mile pass shattering the existing world record of 7.82 seconds at 168mph. " The bike has a 13" diameter GE brushed motor and two Zilla controllers.

Just summited as a starting point.

Yeh we alread though about talking to paul junior to tell him we can help kick his dads ass!

 

[Arlo1]

Your drive system will likely optimize around 800V to 1,000V, yep that's 1 KiloVolt. Drops the amps an order of magnitude from the 100 V bus. This is kill me now AND kill me later territory...

One could design around Hitachi MBN1500E33E2 IGBT modules. Data sheet: http://www.hitachi.co.jp/products/power/pse/images/pdf/igbt/MBN1500E33E2.pdf Those modules are around 5 x 8 inches each! They are rated at 3.3KiloVolts, 1,500 Amps. With a 3,000 Amp surge for 1mS, and a great reverse bias safe operating area. Rise and fall times are 2 uSec with turn off/turn on times around 3 or 4 uSec. So you are likely to run an audible PWM base frequency.

Snubbing and low inductance laminated buss bars are in the mix, as is about 500 hours of computer modeling of the layout. If you stay with 100V you might as well say you are switching a ga-billion-amps/second... handling your switching spikes will take 2 PdD's... and an infi-tacitor or two!

Just thinking of plugging together a 1 KiloVolt LiPo pack is making my hands sweat...

A kilo volt hu?? Thanks big moose.
Ok so I am open minded but while thinking outside the box I can be stuck inside it!
Lets say I have 10 stators with say 18 winding sections each and the magnets dont matter at this point. If I run the 10 sators in series then I can use 1000 volts and wind it for the desired rpm. OR.... I can run 100 volts to each stator and wind it the same! And for that mater have 10 seperat controler and time the stator off just a bit so the battey doesnt take bit amp hits all at once!.
Just a thought how I am thinking about skinning this cat!

 

[Arlo1]

For a race EV, 0000awg would be fine for 5KA.

Full of Yummy mercury!

 

[grindz145]

There is no way you're going to be able to push that kind of current at 100V. BigMoose is right! If there was every an Ebike application for IGBTs this is it. Time to start scouring the locomotive junkyards

It is possible though! Brushless insanity....

 

[Doctorbass]

For a race EV, 0000awg would be fine for 5KA.

Full of Yummy mercury!

Now you guys should begin talking about MCM range cable size.. like 500MCM

And since battery and controller would not be the limit, let's invent a supraconductor multistage axial flux motor!

Could be fun to see you refill your motor of liquid nitrogen before a drag race

Doc

 

[x88x]

If there was every an Ebike application for IGBTs this is it. Time to start scouring the locomotive junkyards

YMMV, but I found quite a few HV/HC IGBTs in a dead Plasma TV I dismantled recently (shattered screen). ..though it sounds like you have enoguh budget for this project that you wouldn't really need to go scavenging.

Could be fun to see you refill your motor of liquid nitrogen before a drag race

Mmmm, that would just be awesome!

 

[Ariane]

Could be fun to see you refill your motor of liquid nitrogen before a drag race

Doc

Nope they just need a Mister Fusion generator with garbage like bananas, soda can, and other thing no plutonium required any more, for this drag electric bike :lol:

Good day!
Black Arrow

 

[bigmoose]

... let's invent a supraconductor multistage axial flux motor!

Doc, you don't really have to invent it. Just give Larry Long a call at LEI. He has designs for a 400 HP (steady state not peak) high temperature (Liquid N2 temps) superconductor 3 phase motor the size of a 3 pound coffee can, and a 50,000 HP motor the size of a garbage can! Here is a little write up on his achievements. He is in Pennsylvania. http://www.mdatechnology.net/techprofile.aspx?id=568 Paper that includes LEI's concept http://gltrs.grc.nasa.gov/reports/2005/TM-2005-213800.pdf

You want to use Y- Ba-Cu-O tape for your high temp superconductor drag motor that will run in liquid nitrogen. Details here http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6TWR-43B8JSB-8&_user=10&_coverDate=02%2F28%2F2001&_rdoc=1&_fmt=high&_orig=search&_origin=search&_sort=d&_docanchor=&view=c&_searchStrId=1544725703&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=b92798dc77fe38733915663c380ffe8d&searchtype=a but you have to buy the paper.

Larry is a nice guy and a pilot. Hands on type. Wouldn't be surprised if he would be interested in a racing application; but he has been on the end of government contracts for a long time, so the price may be very high. The technology is there it is perhaps just at 100X the price of a top fuel bike cost...

 

[Arlo1]

Thanks bigmosse. This is the kind of thing I need to here. Even if I don't get a 3 pound coffee can type motor built It sure makes me try harder to invent one that will try to compeat!

 

[bigmoose]

Your welcome Arlo. I always felt $'s buy development; but for "breakthroughs" you need more creative "heads" thinking and doodling, then drawing and cutting metal... don't ever stop thinking, and the rest will follow!

 

[Kingfish]

Taking the bait, let’s toy with the idea that massive current won’t be an issue, however we resolve it. I am going to borrow a component from Luke’s FET resource page for MEV/HEV applications page.

• Using http://ixdev.ixys.com/DataSheet/VMO1600-02P.pdf just for discussion.
  200 V, 1600 Amp. Downrated 75% = 1200 A
  I = P / V -> 745700 / 200 = 3728.5 Amps
  I/phase = 3728.5 / √3 = 2153 A / phase
  Number of devices = 2153 A / 1200 A = 1.8 units / leg ≈ 2 / leg. Total = 2 * 2 * 3 = 12 devices total.

The advantages are that scary voltage is reduced; although still a safety issue, it is a lot less than 1000 V. With the device we need not worry about wasteful low-power issues.

Motor Modularity:
If we used LN2 and were able to achieve a 10X increase in current density, then an 9-rotor/8-stator AF in parallel would pull the equivalent 27 A (or 270 A at room temp) per stator.

There is also the option of using a smaller diameter rotor which will spin faster and quicker and provide higher utility in trade for lower torque. To offset the lower torque one could gang two or three motors together on a shared ring gear which is the drive shaft connected to the rear wheel via chain.

This idea is not foreign to me as I was trained to operate high-speed turbines in my youth: The ship’s screw is driven by a long shaft connected to a bull gear surrounded by multiples of engines. Massive torque, and very quick on speed changes. Smaller motors = more compaction, modularity, and utilization of space.

Imagine speeds approaching 200 mph using one motor. Approaching 300, slap on another motor. 400? If the motor dies you have the opportunity to swap it out like you could swap out a piston. At high current the chance is the lifetime of these devices will be limited. Multiple motors = survivability; failure or weakness of one doesn’t bring down the system. Lastly, it’s a lot less expensive to build many smaller motors.

I think that would be my approach if I were to build a factor designed to set records, and then keep them.
~KF
PS – thanks Bigmoose for sharing; those are inspiring technical reports!

 

[Arlo1]

Taking the bait, let’s toy with the idea that massive current won’t be an issue, however we resolve it. I am going to borrow a component from Luke’s FET resource page for MEV/HEV applications page.

• Using http://ixdev.ixys.com/DataSheet/VMO1600-02P.pdf just for discussion.
  200 V, 1600 Amp. Downrated 75% = 1200 A
  I = P / V -> 745700 / 200 = 3728.5 Amps
  I/phase = 3728.5 / √3 = 2153 A / phase
  Number of devices = 2153 A / 1200 A = 1.8 units / leg ≈ 2 / leg. Total = 2 * 2 * 3 = 12 devices total.

The advantages are that scary voltage is reduced; although still a safety issue, it is a lot less than 1000 V. With the device we need not worry about wasteful low-power issues.

Motor Modularity:
If we used LN2 and were able to achieve a 10X increase in current density, then an 9-rotor/8-stator AF in parallel would pull the equivalent 27 A (or 270 A at room temp) per stator.

There is also the option of using a smaller diameter rotor which will spin faster and quicker and provide higher utility in trade for lower torque. To offset the lower torque one could gang two or three motors together on a shared ring gear which is the drive shaft connected to the rear wheel via chain.

This idea is not foreign to me as I was trained to operate high-speed turbines in my youth: The ship’s screw is driven by a long shaft connected to a bull gear surrounded by multiples of engines. Massive torque, and very quick on speed changes. Smaller motors = more compaction, modularity, and utilization of space.

Imagine speeds approaching 200 mph using one motor. Approaching 300, slap on another motor. 400? If the motor dies you have the opportunity to swap it out like you could swap out a piston. At high current the chance is the lifetime of these devices will be limited. Multiple motors = survivability; failure or weakness of one doesn’t bring down the system. Lastly, it’s a lot less expensive to build many smaller motors.

I think that would be my approach if I were to build a factor designed to set records, and then keep them.
~KF
PS – thanks Bigmoose for sharing; those are inspiring technical reports!

There you go, Just need to visulize it. So is there a way we can graph the gains in torque vs rotor/motor dia? Then we can pick a point on the graph and use it to start then see how many stator sections you can fin in the width of the bike!

 

[Kingfish]

So is there a way we can graph the gains in torque vs rotor/motor dia? Then we can pick a point on the graph and use it to start then see how many stator sections you can fin in the width of the bike!

That would take some iterative work, and should be explored. In the meantime allow me to press on by means of an idea.

With FEMM, I modeled numerous configurations using thick and thin magnets and determined that fat magnets are not as useful as thinner magnets in greater quantity.

Conversely a single motor would be huge: The number of stators would be limited with thick magnets. Let's flip it around and look at it another way: Smaller motors, more stators in parallel, spread the load out between multiple motors.

• 18 inches = Available motor width. Subtract an inch for each end-plate = 16 inches.
  Converted to metric: 16 * 25.4 = 406.4 mm
  Formula for number of rotor and stators: ((Mw + AG) * Pairs) + Mw
  Where
  Mw = Magnet width
  AG = air gap between magnet faces (includes stator)
  Pairs = Rotor-Stator pairs
  Let’s say we want to use 12 mm thick magnets and 12 mm wide air gaps.
  406.4 mm available width – 12 mm for end-rotor = 394.4.
  394.4 / (Mw + AG) => 398.4 / 24 = 16.43 pairs ≈ 16 pairs.

That’s a large sandwich but not unreasonable if you’ve imagined how many rotors are in jet, gas, and steam turbines. It could present a current-balancing issue though.

We want these motors designed to rotate at 4,000 rpm minimum without any issues, therefore the OD of the rotor all of a sudden begins to look like an 8-inch hub motor and less like a 14-inch windmill turbine. Next, we use two of these motors – in parallel mated with double-helical gearing to a bull gear that has the sprocket & chain. This assembly can be integrated into the transmission area, perhaps with clutching. The transmission is oil-cooled. 8)

• Two motors, 16 stators each serving up 2153 Amps of current / phase (if 3-P) = 2153 / 32 = 67.3 A / stator at room temperature. That looks doable.

The $64,000 question is: Will it work

• The target Torque will be 2189 Nm / 32 = 68.4 nm / stator spinning at 3253 rpm (31-inch tire at 300 mph).

Let’s compare that to my Plan-D design in Doing the Math:

• Using 4 mm tall magnets with a 3-rotor/2-stator motor at more or less the same rotor OD spinning at 420 rpm would produce > 33.9 Nm => 33.9 / 2 = 16.95 Nm / stator @ 30 mph. In conclusion, my little AF hub motor can theoretically put out 25% of the required torque at 10% of the speed. Admittedly we are comparing apples to oranges in terms of purposing – though this comparison exposes that we are on the correct track and quite possibly have a monster-configuration for success.

I can tell you that 12 mm tall magnets will come with quite a large Tesla rating. Perhaps we can trade that in exchange for heat-tolerance which I believe we sorely need to consider.

~KF

 

[fechter]

Even if you use plain old copper for the motor windings, it might be some benefit to chill the windings with LN2 before a run to decrease the resistance and make it take a little longer to overheat. I'm not sure if that's practical, as you'd need to keep the bearings warm, but it might make a considerable difference. Plus it would look cool if there was a cloud of fog rolling out of the motor compartment.

Really big induction motors can be found surplus for cheap. Big VFAC drives can be found surplus too, but they are BIG!

 

[liveforphysics]

Current in a controller does not stack linearly, aka, you can't just scale-up. (it only does on paper)


This is why Moose suggested higher voltages.

It's fairly common to find 1200v 1200amp IGBT modules on flea-bay for a few hundred bucks. They switch slowly, so they won't be able to have effective current control on low inductance motors, but they should be able to get the job done for a drag bike with a motor built around working with them.

 

[bigmoose]

Even if you use plain old copper for the motor windings, it might be some benefit to chill the windings with LN2 before a run to decrease the resistance and make it take a little longer to overheat. I'm not sure if that's practical

You want to look at figure 10 11 and 12 in the reference paper I posted above. That is exactly what was tested, and the data is in the paper... out in the open, free to just read! After you digest it, well it blows your mind, in a good way. Look seriously at the drop in copper resistivity at LN2 temperatures... there is a nugget there.

 

[Kingfish]

¡Ay, caramba! I forgot to calculate Force:

• F = IL x B
  Ï„ = 2rF = 2rBIL
  B = 0.5 T (this value could be higher depending on the magnets and air gap)
  r = 15.5 inches / 0.3937 meters
  I = 360.64 A
  Ï„ = 2189 Nm

Plan-D for comparison:

• r = 0.090 meters for Plan-D, therefore
  if Ï„ = r x F, r = 0.3937 m, and Ï„ = 2189 Nm, then
  F = Ï„ / r = 2189 / 0.3937 = 5560 N

Calculate new Ï„:

• r1 x F1 = r2 x F2 => F2 = (r1 x F1) / r2
  0.3937 x 5560 = 0.090 x F2 => F2 = (0.3937 x 5560) / 0.090 = 24321 N
  
• Therefore the amount of Force (F) for Plan-D is 377 N, or 188.5 N / stator.
• Proposed AF motor Force would be 24321 N, or 24321 / 32 = 760 N / stator.

Again – the proposed motor has 4X the Force as the Plan-D motor.

Luke, what are the typical voltages that EV cars & trucks run at? It seems to me that one of these is using 288 V. Why not use similar parts?

~KF

 

[Arlo1]

Current in a controller does not stack linearly, aka, you can't just scale-up. (it only does on paper)


This is why Moose suggested higher voltages.

It's fairly common to find 1200v 1200amp IGBT modules on flea-bay for a few hundred bucks. They switch slowly, so they won't be able to have effective current control on low inductance motors, but they should be able to get the job done for a drag bike with a motor built around working with them.

Im thinking multiple stators and multiple controlers and if needed keep the battery packs seperate but this is why im talking about it first so we can come up with some the best guesses to make it work!